Cabling for Fluke 732A/B, etc.

[michthai]

Definetely check the switch first. It's almost certain to be your problem. Try another channel if you have to. You can write a routine to exercise those switches that may clear up your problem.

[Hawaii596]

I did stop the program and tried manually exercising the switch.  It was weird that it happened at exactly the reading set after I changed some connections.  I didn't touch the HI or LO connections at all on any cell.  I just un-jumpered GUARD from chassis ground on all of the DC standards.  I did some repetitive opens/closes on that switch while monitoring the 3458A uV reading.  And it stayed consistent and relatively stable.  What I might do is manually measure through resistance through those contacts end-to-end on each cable and see if there is a significant difference on that one.  And I think its noteworthy that it is 10 VDC potential on each line.  So it is 10VDC versus 10VDC in the 100 mVDC range.  I had a nice normal drift curve in my relative readings (on an Excel chart).  And I am plotting all of the readings.  That one delta made an instantaneous shift down at that time point and continued it's normal drift curve.  So it doesn't "LOOK" like a flaky switch contact.  But as a long time technician, yes, I need to prove that one way or the other.  Actually, the HP 3488A switch matrix uses a modular card.  The interface with the wires connected to is (with screw down contacts) is one piece, and that plugs into the main card (which has all the relays).  Maybe I could try screwing down all the contacts, reseating the card, then if I have another of that model card (which I think I do), I could try swapping out the card.  That should give me some pretty good probability of answering it.

Thanks for the input.

[michthai]

I'm not too sure if using the HP 3488A is your best option. The Guildline low thermal quad scanners (6664 series) are generally the best devices for these types of measurements.

[Hawaii596]

I know it isn't the best hardware "yet."  I'd like to get a Dataproof, which I'm told is also very good.  This has been a shoestring budget project.  Maybe when it is all done I'll post some details.  I have primary resistance (Guildline 9975, Thomas 1 Ohm, SR104 10K, high current module and various adapters), AC voltst (Fluke 792A with all the trimmings - shunts, etc.; not to mention a 540B that I may use as a secondary along with some other Datron hardware), and of course my nine bank DC referenc set that I hope to eventually bring down to sub ppm uncertainty).  I got a nice  hand-me-down copper mesh screen room and line filters.  But for now, still getting up and running.  As far as the 3488A goes, I think it may be a factor of which model card, and what kind of shape the card is in.  I have a stack of spare cards that are not all the same model.  I may shop through them and see the low thermal specs, and possibly even buy a used lower thermal emf card (for now).  But I do agree, the lower the thermals, the better.  So far, by using moving average, I have about 20 readings per day per cell, and over about 10 days or so, I have been tracking to around 0.1 ppm predictability.  I am seeing a little daily shifting late in the afternoon (which I attribute to Texas heat).  So we are shopping for a Liebert box to put in there, and eventually we will add an airlock to control RH as well.  So, as I said, it is still in project phase.

But good input on the low thermal scanner.  I'm not particularly familiar with the Guildline.  I am assuming it is similar to the Dataproof 160B.  Back to it.  I have some Baratrons to calibrate and "real" work, before I can go back to my project.  Lots of cool projects; just not enough time to get them all done.  Thanks for the inputs.

P.S. - For all readers.  Still soliciting resumes for our position.  We really need someone to take over this project (if I may put in a shameless plug); along with good broad DC/LF and RF/Microwave cal skills.

[jwilley127]

For low thermal wire and accessories for precision resistance, voltage or thermometry measurements check out: http://www.jswilley.com/Accessories.html